Rahim Panahi - Dept. Electrical Engineering Sharif University Tehran
Iman Gholampour - Dept. Electrical Engineering Sharif University Tehran
Mansour Jamzad - Dept. Computer Science Sharif University Tehran

Tracking objects using Mean Shift algorithm failswhen there is a full/partial occlusion or when the backgroundcolor and the desired object are close. In this paper we proposeda method using Kalman Filter and Mean Shift for handling thesesituations. Using similarity measure of Mean Shift algorithm weare able to detect an occlusion. Kalman Filter comes into the playfor occlusion handling in a Buffer-Mode Process. Weimplemented this algorithm both on PC and DSP 64x+ TexasInstrument and the results are both tabulated. The results revealthe ability of our method to locate the object soon after occlusiondisappearance.

Occlusion handling; MeanShift ; Kalman Filter Tracking ; DSP Processors ; Prediction